Professional Documents
Culture Documents
studies, and process for requesting access can be observed in the per protocol analysis and at Week 4. Overall, 12.1% of Mg–vitamin B6
found at: https://www.clinicalstudydatarequest. treated and 17.4% of Mg-treated subjects experienced AEs potentially treatment related.
com
Methods
Trial design
This was an 8-week, Phase IV, randomized, controlled, investigator-blinded, parallel-group trial
stratified by sex (EudraCT Number: 2015-003749-24) (Fig 1). The trial was carried out at 4 clini-
cal trial centers in France. Healthy subjects completed a pre-trial telephone interview within 1
week prior to screening; the baseline visit took place <2 weeks after the screening visit (Fig 2).
The 8-week treatment period comprised visits at Week 4 and Week 8. Subjects were randomized
Fig 2. Trial design. � Four subjects did not proceed to the treatment phase and did not receive any investigational product due to major protocol
deviations. Mg, magnesium.
https://doi.org/10.1371/journal.pone.0208454.g002
1:1 to treatment with either the magnesium–vitamin B6 (Mg–vitamin B6) combination or mag-
nesium (Mg) alone. Subjects were randomly divided at baseline in order to avoid systematic dif-
ferences with respect to known or unknown variables that could affect outcomes.
Subjects
Subjects were between 18 and 50 years of age with moderate to extremely severe stress at
screening, defined as a DASS-42 stress subscale score of >18 [20]. Additionally, subjects must
have presented with suboptimal serum magnesium concentrations (assessed locally at each of
the trial sites), defined as serum magnesium concentrations between 0.45 mmol/L and 0.85
mmol/L; these were measured locally at the investigation centers. The upper limit of the serum
magnesium cut-off was chosen based on previous work establishing an evidence-based refer-
ence interval (central 95th percentile) as 0.75 mmol/L to 0.95 mmol/L with a mean concentra-
tion of 0.85 mmol/L. The serum magnesium cut-off of 0.85 mmol/L has been previously
determined as the lower limit adjusted to a value for health in a trial of over 15,000 subjects
[21]. At screening, magnesium concentrations were assessed with blood serum samples. Dur-
ing the study (baseline, Week 4 and Week 8), magnesium concentrations were assessed in the
erythrocytes; erythrocytes are thought to accurately represent magnesium cell content within
the body and provide an accurate reflection of the whole body magnesium status. Additional
inclusion criteria were, a body mass index (BMI) of >18.5 and �29.9 kg/m2, and the use of an
effective method of contraception during the trial period for female subjects. Key exclusion cri-
teria included: exposure to therapies prohibited by the protocol (including levodopa, quini-
dine, and proton-pump inhibitors) within 3 months prior to screening; concomitant
conditions or diseases that could make subjects non-evaluable for the primary endpoint; severe
hypomagnesemia (defined as serum magnesium of �0.45 mmol/L) [21]; participant-reported
moderate or severe kidney failure; confirmed diagnosis of type 1 or 2 diabetes; any known
addiction to drugs and alcohol; alcohol intake of �3 drinks per day.
Interventions
Each participant received coated tablets of either Mg–vitamin B6 combination (Magne B6; 470
mg magnesium lactate dihydrate and 5 mg pyridoxine hydrochloride) or Mg alone (Magne-
spasmyl [Mg]; 465.4 mg magnesium lactate dihydrate). Tablets were self-administered orally,
with subjects taking 6 coated tablets per day (corresponding to approximately 300 mg elemen-
tal magnesium with or without 30 mg vitamin B6) divided into 3 intakes (2 tablets during each
meal [breakfast, lunch, and dinner]) over a period of 8 weeks.
Tablets were provided in treatment sets with 50 tablets, dispatched in 5 blister packs of 10
tablets each. Subjects received 4 treatment kits at randomization for the following 4 weeks,
after which they received an additional 4 treatment kits at the Week 4 visit for use until the
end of the trial.
Objectives
The primary objective was to compare magnesium in combination with vitamin B6 supple-
mentation versus magnesium alone on stress, evaluated by the stress subscale from the DASS-
42 test, in healthy adults with stress and suboptimal serum magnesium concentrations. A sec-
ondary objective was to evaluate the safety profile of the Mg–vitamin B6 combination versus
Mg alone, as determined by the incidence of adverse events (AEs).
Sample size
An overall sample size of 268 (134 in each treatment arm) was selected to obtain 119 evaluable
subjects per treatment arm (assuming 10% of subjects would be non-evaluable), and achieve
80% power to detect a difference of 3 points on the DASS-42 stress subscale score change from
baseline at Week 4 and Week 8 between intervention arms (delta of superiority corresponding
to ~15% of baseline value, which was expected to be ~20). Group standard deviations (SD) of 8
points were estimated with a significance level (alpha) of 0.05 using a two-sided equal-variance
t-test and a Mann–Whitney test, and assuming normal distribution.
The clinically relevant difference of 3 points was defined with expert clinicians. The SD esti-
mation was based on the evaluation of DASS-42 stress subscale score of “emailed standardized
cognitive behavioral treatment of work-related stress” [22].
Statistical methods
The modified Intention-To-Treat (mITT) population was defined as all subjects with an evalu-
able DASS-42 stress subscale score at baseline and at least one other time point during the
treatment period. The Per Protocol (PP) population was defined as all subjects included in the
mITT analysis without any major protocol deviations.
Of note, the mITT population comprised some subjects who had a baseline DASS-42
stress subscale score �18. This can be attributed to variations in DASS-42 stress subscale
score between screening and baseline (~two weeks occurred between screening and base-
line). Therefore, the statistical analysis plan was amended after the database lock based on a
decision from the scientific committee, such that the PP population should include all sub-
jects in the mITT without any major protocol violations, and with a DASS-42 stress subscale
score >18 at baseline.
The population included in the Safety Set (SS) comprised all subjects included in the trial
with at least one consumption of trial product and was used for safety evaluations.
Differences between treatment arms in change from baseline to Week 4 or Week 8 in
DASS-42 stress subscale scale score were calculated by means of a repeated measures analysis
of covariance (ANCOVA), with DASS-42 stress subscale score at baseline and stratification
factor (sex) as covariates, and visit and interaction between visit and treatment group and
interaction between baseline value and treatment group as fixed effects.
Adjusted mean was calculated from model ANCOVA adjusted by baseline value of DASS-
42 and the interaction of baseline and treatment.
Subgroup analyses
Based on European Medical Agency (EMA) guidelines for the adjustment of baseline covari-
ates [23], the statistical analysis plan was amended after the blinded database lock to include
subgroup analyses of primary and secondary efficacy endpoints by baseline stress level, as a
significant interaction (p < 0.05) was present between baseline DASS-42 stress subscale scores
and treatment (assessed firstly as a continuous variable and then categorically by class [‘normal
and moderate stress’ versus ‘severe and extremely severe stress’]).
Results
Participant disposition
Between May 2016 and January 2017, 854 subjects were screened and 268 were enrolled (134
in each treatment arm) (Fig 1). Two subjects in each treatment arm received no trial treat-
ment, and were excluded from the mITT and PP analysis. Of note, 26 subjects (13 in each
arm) in the mITT analysis had baseline DASS-42 stress subscale scores �18 despite the inclu-
sion criteria used at screening. This was due to variation in stress levels between screening and
baseline. These 26 subjects were excluded from the PP analysis in line with statistical analysis
plan, in addition to a further five subjects (two in the Mg–vitamin B6 combination arm and
three in the Mg alone arm) who were excluded due to an observed treatment adherence of
�75%, defined as a major protocol deviation. For the subgroup analyses specified in the
results, the 26 subjects with baseline DASS-42 stress subscale scores �18 were included in the
mITT but excluded from the PP analysis. Overall, 260 subjects (130 in each arm) completed
the trial (Fig 1).
https://doi.org/10.1371/journal.pone.0208454.t001
Table 2. Distribution of DASS-42 stress subscale scores at baseline and Week 8 (mITT population).
Baseline Week 8
DASS-42 stress score Mg–vit B6 combination Mg Total Mg–vit B6 combination Mg Total
(N = 132) (N = 132) (N = 264) (N = 132) (N = 132) (N = 264)
Normal (scorea 0–14), n (%) 4 (3.0) 6 (4.5) 10 (3.8) 75 (57.3) 69 (53.1) 144 (55.2)
Mild (scorea 15–18), n (%) 9 (6.8) 7 (5.3) 16 (6.1) 18 (13.7) 21 (16.2) 39 (14.9)
Moderate (scorea 19–25), n (%) 41 (31.1) 35 (26.5) 76 (28.8) 29 (22.1) 17 (13.1) 46 (17.6)
Severe (scorea 26–33), n (%) 46 (34.8) 59 (44.7) 105 (39.8) 7 (5.3) 15 (11.5) 22 (8.4)
Extremely severe (scorea 34–42), 32 (24.2) 25 (18.9) 57 (21.6) 2 (1.5) 8 (6.2) 10 (3.8)
n (%)
a
Stress subscale score
DASS, Depression Anxiety Stress Scale; mITT, modified intention-to-treat; Mg, magnesium; Vit B6, vitamin B6.
https://doi.org/10.1371/journal.pone.0208454.t002
overweight (25–<30 kg/m2) status (mean [SD] DASS-42 for those not overweight versus over-
weight: 27.9 [7.1] versus 26.9 [7.0]; p > 0.05).
Overall, 39% of subjects in the Mg–vitamin B6 combination arm and 40% in the Mg arm
reported at least one medication prior to trial entry, 39% in both arms had ongoing medication
at baseline, and 25% and 21%, respectively, started at least one medication between baseline
and Week 8. None of the prior medications have any known interaction with the trial product
or effect on stress.
Treatment adherence
Between baseline randomization and Week 8, mean (SD) assumed treatment adherence was
94% (7.1) in the Mg–vitamin B6 combination arm and 93.0% (9.2) in the Mg arm; 96% of sub-
jects (127/132) in each arm had an assumed treatment adherence of �80%.
Efficacy
Overall change from baseline in DASS-42 stress subscale score. In the mITT popula-
tion, both treatments reduced DASS-42 stress subscale score from baseline to Week 8, reduc-
ing the overall proportion of subjects with severe or extremely severe stress at baseline from
approximately 60% to approximately 12% (Table 2). The DASS-42 stress subscale score
improved by 44.9% from a mean (SD) of 27.7 (7.3) points at baseline to 14.5 (7.4) points at
Week 8 in the Mg–vitamin B6 combination arm, representing a change in adjusted mean of
−12.44 points (95% confidence interval [CI] −13.83 to −11.05). Scores also improved in the
Mg group, by 42.4%, from 27.6 (7.0) points at baseline to 15.3 (9.5) points at Week 8, with a
change in adjusted mean of −11.72 points (95% CI −13.10 to −10.33). The difference between
treatment arms was not statistically significant (0.72 points, 95% CI −1.15 to 2.59, p > 0.05)
(Table 3). Similar findings were observed in the PP population with no statistically significant
difference between treatment arms (1.06 points, 95% CI −0.99 to 3.10, p > 0.05 (Table 3).
In the mITT population, both treatment arms improved DASS-stress subscale score from
baseline to Week 4 (Mg–vitamin B6, −8.94 points, 95% CI −10.22 to −7.65; Mg, −7.58, 95% CI
−8.86 to −6.30); the difference between treatment arms was not statistically significant (1.35
points, 95% CI −0.36 to 3.06, p > 0.05) (Table 3). Similar improvements from baseline to
Week 4 were observed in the PP population, with no statistically significant difference between
treatment arms (1.55 points, 95% CI −0.33 to 3.43, p > 0.05; Table 3).
Change in DASS-42 stress subscale score in baseline stress severity subgroups. Accord-
ing to recent EMA guidelines, the identification of heterogeneous benefits within samples
Table 3. Change in DASS-42 stress subscale score from baseline to Week 4 and to Week 8.
mITT population Mg–vit B6 combination (N = 132) Mg (N = 132)
Change from baseline to Week 4a (95% CI) −8.94 (−10.22 to −7.65) −7.58 (−8.86 to −6.30)
Difference between treatment arms 1.35 (−0.36 to 3.06), p = 0.1203
Change from baseline to Week 8a (95% CI) −12.44 (−13.83 to −11.05) −11.72 (−13.10 to −10.33)
Difference between treatment arms 0.72 (−1.15 to 2.59), p = 0.4472
PP population Mg–vit B6 combination (N = 117) Mg (N = 116)
Change from baseline to Week 4ab (95% CI) −9.59 (−11.03 to −8.15) −8.04 (−9.45 to −6.63)
Difference between treatment arms 1.55 (−0.33 to 3.43), p = 0.1056
Change from baseline to Week 8ab (95% CI) −13.26 (−14.81 to −11.71) −12.21 (−13.73 to −10.68)
Difference between treatment arms 1.06 (−0.99 to 3.10), p = 0.3095
a
Difference from baseline in adjusted mean.
b
Subjects with subscale scores �18 baseline were excluded from the PP population.
CI, confidence interval; DASS, Depression Anxiety Stress Scale; mITT, modified intention-to-treat; Mg, magnesium;
PP, per protocol; Vit B6, vitamin B6.
https://doi.org/10.1371/journal.pone.0208454.t003
warrants additional subgroup analyses [23], thus a subgroup analysis was performed, in line
with the statistical analysis plan. The interaction test to assess association between DASS-42
stress subscale score at baseline and treatment showed a statistically significant association in
both the mITT population (p = 0.0097) and the PP population (p = 0.0171). Therefore, an
analysis of change in DASS-42 stress subscale score by subgroup according to baseline score
was carried out in both the mITT and PP populations.
For mITT population, the subgroups were defined as normal to moderate stress and severe
to extremely severe stress; in the PP population the subgroups were defined as moderate stress
and extreme to extremely severe stress. The severe to extremely severe stress subgroup com-
prised subjects with baseline DASS-42 stress subscale scores between 26 and 42 (mITT and
PP); the normal to moderate stress subgroup comprised subjects with baseline stress subscale
scores between 0 and 25 (mITT) [20]; the moderate stress subgroup comprised subjects with
scores between 19 and 25 and did not include subjects with baseline DASS-42 stress subscale
scores �18, who were already excluded from the PP as per the statistical analysis plan. For
both mITT and PP population analyses, baseline DASS-42 stress subscale scores were similar
between treatment arms in the severe and extremely severe stress subgroup (mITT: Mg–vita-
min B6 combination, n = 78; Mg, n = 84 and PP: Mg–vitamin B6 combination, n = 76; Mg
n = 84) (Fig 3) and the normal to moderate (mITT population) and moderate (PP population)
stress subgroups (mITT: Mg–vitamin B6 combination, n = 54; Mg, n = 48 and PP; Mg–vitamin
B6 combination n = 41; Mg, n = 32) (Fig 3).
In the mITT population, both treatment arms in the severe and extremely severe stress sub-
group had an improvement in DASS-42 subscale score from baseline to Week 8. The Mg–vita-
min B6 arm had a 50.1% improvement (−16.36 points, 95% CI, −18.27 to −14.44) and the Mg
arm had a 41.3% improvement (−13.20 points, 95% CI, −15.05 to −11.36). There was a statisti-
cally significant improvement from baseline to Week 8 in Mg–vitamin B6-treated subjects
compared with the Mg group (Fig 3A). The improvement was 23.9% greater (3.16 points, 95%
CI 0.50 to 5.82, p = 0.0203) for the Mg–vitamin B6 combination group versus the Mg group.
There was no significant difference between treatment arms in subjects with normal to moder-
ate stress (−2.36 points, 95% CI −4.99 to 0.27, p > 0.05) (Fig 3A).
Similar results were observed in the PP population (Fig 3B). In subjects with severe or
extremely severe stress, both treatment arms improved DASS-42 stress subscale score. The
Fig 3. Change in DASS-42 stress subscale score from baseline to Week 8 in the mITT (A) and PP (B) subgroup populations. CI, confidence interval; DASS,
Depression Anxiety Stress Scale; Mg, magnesium; mITT, modified intention-to-treat; PP, per protocol; Vit B6, vitamin B6.
https://doi.org/10.1371/journal.pone.0208454.g003
Mg–vitamin B6 arm had an improvement of 49.4% (−16.09 points, 95% CI −18.02 to −14.16)
and the Mg arm had an improvement of 41.3% (−13.20 points, 95% CI−15.04 to −11.36). A
statistically significant 21.9% (2.89 points, 95% CI 0.22 to 5.56, p = 0.0339) greater improve-
ment from baseline to Week 8 was observed in Mg–vitamin B6-treated versus Mg-treated sub-
jects. There was no significant difference between treatment arms in subjects with moderate
stress (−1.96 points; 95% CI −5.11 to 1.20, p > 0.05).
Changes from baseline to Week 4 are shown in Table 4. In the mITT population, in sub-
jects with severe or extremely severe stress, the improvement from baseline was 38.2% (3.37
points; 95% CI 1.02 to 5.73) greater with Mg–vitamin B6 combination compared with Mg, and
the difference in change between treatments was statistically significant (p = 0.0053). In sub-
jects with normal to moderate stress, the change from baseline was not significantly different
between treatment arms (−1.22 points; 95% CI −3.73 to 1.29; p > 0.05).
Change from baseline to Week 4 in the PP population was consistent with the mITT popu-
lation findings (Table 4). In subjects with severe or extremely severe stress, the improvement
from baseline was 39.6% (3.50 points, 95% CI 1.13 to 5.86) greater with the Mg–vitamin B6
combination compared with Mg, and the difference was statistically significant (p = 0.0041).
In subjects with moderate stress, there was no significant difference between treatment arms
(−1.79 points, 95% CI −4.89 to 1.32; p > 0.05).
Safety
Overall, 41% (54/132) of subjects in the Mg–vitamin B6 combination arm and 40% (53/132) in
the Mg arm experienced at least one AE. In total, 12% (16/132) of subjects in the Mg–vitamin
B6 combination group and 17% (23/132) in the Mg group experienced at least one AE consid-
ered possibly related to trial treatment, as assessed by the trial physician. The most common
treatment-related AE was diarrhea, reported by 4.5% (6/132) of subjects in the Mg–vitamin B6
combination arm and 7.6% (10/132) in the Mg arm (Table 5). One (0.8%) subject in the Mg
arm had at least one AE of severe intensity (gastroenteritis); no such events occurred in the
Mg–vitamin B6 combination arm.
Treatment dose was interrupted for two subjects in the Mg–vitamin B6 combination arm
and four in the Mg arm due to AEs. All six subjects recovered from the AEs. One participant
Table 4. Change in DASS-42 stress subscale score from baseline to Week 4 by DASS-42 subscale subgroup score at baseline.
mITT population Severe to extremely severe stress (N = 162) Normal to moderate stress (N = 102)
Mg–vit B6 combination Mg (N = 84) Mg–vit B6 combination Mg (N = 48)
(N = 78) (N = 54)
Baseline, adjusted mean (SE) 32.67 (0.52) 31.93 (0.47) 20.53 (0.52) 20.18 (0.58)
Week 4, adjusted mean (SE) 20.47 (0.96) 23.10 (0.91) 14.90 (0.87) 13.33 (0.94)
Change in adjusted mean (95% CI) −12.21 (−13.90 to −10.51) −8.83 (−10.47 to −5.63 (−7.35 to −3.91) −6.85 (−8.68 to
−7.20) −5.03)
Difference in change in adjusted mean between arms 3.37 (1.02 to 5.73), p = 0.0053 −1.22 (−3.73 to 1.29), p = 0.3354
(95% CI)
PP population Severe to extremely severe stress (N = 160) Moderate stress (N = 73)
Mg–vit B6 combination Mg (N = 84) Mg–vit B6 combination Mg (N = 32)
(N = 76) (N = 41)
Baseline, adjusted mean (SE) 32.57 (0.53) 31.92 (0.47) 22.13 (0.34) 21.94 (0.40)
Week 4, adjusted mean (SE) 20.25 (0.97) 23.08 (0.91) 15.88 (1.04) 13.90 (1.18)
Change in adjusted mean (95% CI) −12.33 (−14.04 to −10.61) −8.83 (−10.46 to −6.24 (−8.30 to −4.19) −8.03 (−10.36 to
−7.20) −5.70)
Difference in change in adjusted mean between arms 3.50 (1.13 to 5.86), p = 0.0041 −1.79 (−4.89 to 1.32), p = 0.2550
(95% CI)
CI, confidence interval; DASS, Depression Anxiety Stress Scale; mITT, modified intention-to-treat; Mg, magnesium; PP, strict per protocol analysis; SE, standard error;
Vit B6, vitamin B6. Standard error rather than standard deviation was calculated for subgroups to account for the means being adjusted.
https://doi.org/10.1371/journal.pone.0208454.t004
Table 5. Treatment-related adverse events occurring in �1% of the total population (SS).
System Organ Class, Preferred Term Mg–vit B6 combination (N = 132) Mg (N = 132) Total (N = 264)
Subjects with �1 TRAE, TRAEs, Subjects with �1 TRAE, TRAEs, Subjects with �1 TRAE, TRAEs,
n (%) n (%) n (%) n (%) n (%) n (%)
Total 16 (12.12) 26 (100) 23 (17.42) 43 (100) 39 (14.77) 69 (100)
Gastrointestinal disorders 11 (8.33) 18 18 (13.64) 36 29 (10.98) 54
(69.23) (83.72) (78.26)
Abdominal discomfort 1 (0.76) 1 (3.85) 0 (0.00) 0 (0.00) 1 (0.38) 1 (1.45)
Abdominal distension 2 (1.52) 2 (7.69) 0 (0.00) 0 (0.00) 2 (0.76) 2 (2.90)
Abdominal pain 3 (2.27) 3 (11.54) 6 (4.55) 10 9 (3.41) 13
(23.26) (18.84)
Abdominal pain upper 0 (0.00) 0 (0.00) 2 (1.52) 2 (4.65) 2 (0.76) 2 (2.90)
Constipation 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
Diarrhea 6 (4.55) 6 (23.08) 10 (7.58) 19 16 (6.06) 25
(44.19) (36.23)
Dry mouth 1 (0.76) 1 (3.85) 0 (0.00) 0 (0.00) 1 (0.38) 1 (1.45)
Dysphagia 1 (0.76) 1 (3.85) 0 (0.00) 0 (0.00) 1 (0.38) 1 (1.45)
Feces soft 3 (2.27) 3 (11.54) 2 (1.52) 2 (4.65) 5 (1.89) 5 (7.25)
Flatulence 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
Frequent bowel movements 1 (0.76) 1 (3.85) 0 (0.00) 0 (0.00) 1 (0.38) 1 (1.45)
Nausea 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
General disorders and administration site 1 (0.76) 1 (3.85) 0 (0.00) 0 (0.00) 1 (0.38) 1 (1.45)
conditions
Asthenia 1 (0.76) 1 (3.85) 0 (0.00) 0 (0.00) 1 (0.38) 1 (1.45)
Metabolism and nutrition disorders 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
Increased appetite 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
Nervous system disorders 2 (1.52) 2 (7.69) 5 (3.79) 5 (11.63) 7 (2.65) 7 (10.14)
Dizziness 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
Headache 2 (1.52) 2 (7.69) 3 (2.27) 3 (6.98) 5 (1.89) 5 (7.25)
Hypersomnia 0 (0.00) 0 (0.00) 1 (0.76) 1 (2.33) 1 (0.38) 1 (1.45)
Psychiatric disorders 2 (1.52) 2 (7.69) 0 (0.00) 0 (0.00) 2 (0.76) 2 (2.90)
Sleep disorder 2 (1.52) 2 (7.69) 0 (0.00) 0 (0.00) 2 (0.76) 2 (2.90)
Skin and subcutaneous tissue disorders 3 (2.27) 3 (11.54) 1 (0.76) 1 (2.33) 4 (1.52) 4 (5.80)
Dermatitis acneiform 1 (0.76) 1 (3.85) 1 (0.76) 1 (2.33) 2 (0.76) 2 (2.90)
Rash papular 2 (1.52) 2 (7.69) 0 (0.00) 0 (0.00) 2 (0.76) 2 (2.90)
Mg, magnesium; SS, safety set; TRAE, treatment-related adverse event; Vit B6, vitamin B6.
https://doi.org/10.1371/journal.pone.0208454.t005
in the Mg arm experienced a SAE (severe gastroenteritis that led to hospitalization), and was
subsequently withdrawn from the trial. This SAE was considered not to be related to trial med-
ication as assessed by the trial physician; therefore, the participant received the treatment until
trial withdrawal. No deaths occurred during the course of the trial.
Discussion
This is the first randomized clinical trial to evaluate the effects of magnesium and vitamin B6
(Mg–vitamin B6) supplementation versus magnesium alone (Mg) on stress in individuals with
stress and low serum magnesium concentrations using a validated psychometric measure of
perceived stress (the DASS-42 stress subscale). Both treatments were administered in accor-
dance with the recommended posology, relevant for the prevention and treatment of magne-
sium deficiency and associated symptoms, including stress-related symptoms like mild
anxiety, and nervousness. The 300 mg elemental magnesium provided by the treatments cov-
ers 75–100% of the recommended daily intake [24], and a dosing regimen spread over two-to-
three intakes per day is generally recommended to improve magnesium bioavailability [25].
Formulated as magnesium salt of organic acid (lactate), both Magne B6 and Magnespasmyl
have greater solubility than magnesium salts of inorganic acids, and are therefore associated
with better absorption and increased magnesium bioavailability [26–29]. Although previous
results from experimental clinical trials were suggestive of a beneficial effect of the specific
combination of magnesium lactate and vitamin B6 in a 10:1 ratio on subjective mood follow-
ing several weeks of treatment [17], no randomized clinical trial has previously investigated
the efficacy of such a formula on perceived stress in healthy adults.
Both interventions rapidly reduced stress from baseline, as indicated with a reduction in
DASS-42 stress subscale score across both treatment arms of approximately 30% (~8 points) at
Week 4 and 40% (~12 points) at Week 8 in the whole trial population (Table 3). Given that
the mean (SD) DASS-42 stress subscale score across all subjects at baseline was 27.7 (7.1), this
represents a clinically relevant reduction, being sufficient to move a participant with severe
stress to the moderate category and a participant with moderate stress to the mild category. At
Week 8, the stress reduction with Mg–vitamin B6 treatment was approaching 50% compared
with baseline in severely stressed individuals. Numerous studies have confirmed the robust
psychometric properties of the DASS-42 test in both adult and elderly populations with mood
or anxiety disorders, with internal consistency ranging from 0.88 to 0.95 for the stress subscale
[19, 30, 31]. The large magnitude of effect observed in this trial provides strong evidence for
the benefits of Mg supplementation in stressed individuals with low serum magnesium; how-
ever, a placebo-controlled trial would be required to determine the true extent of benefits,
especially given the large placebo effects observed in previous studies [17].
The benefits of magnesium in this population with low serum magensium concentration
could be attributed to the effect of Mg levels on resistance to stress. Catecholamines and corti-
costeroids released during periods of stress decrease serum magnesium concentration through
urinary excretion [3,5]. Conversely, low serum magnesium concentration increases the release
of these same stress-associated hormones, leading to a positive feedback loop that enhances
both the release of stress hormones and the depletion of magnesium [4,6].
Although superiority of Mg–vitamin B6 versus Mg was not demonstrated in the whole pop-
ulation, a statistically significant interaction was identified between baseline DASS-42 stress
subscale score and treatment. According to recent regulatory guidance and EMA guidelines,
the identification of a qualitative interaction that illustrates heterogeneous benefits within trial
populations warrants further subgroup analyses [23, 32], thus a subgroup analysis was per-
formed in line with the statistical analysis plan. Although a possible bias should be considered,
due to the fact that randomization was performed on the initial whole population and not on
the subgroups, this analysis revealed significantly greater reductions in the symptoms of stress
with Mg–vitamin B6 than with Mg alone in subjects with severe and extremely severe stress,
but no difference in those with normal to moderate stress. The majority of the improvements
in DASS-42 stress subscale score occurred between baseline and Week 4 for both treatment
arms, with incremental improvements seen between Week 4 and Week 8. Of note, in subjects
with severe or extremely severe stress, Mg–vitamin B6 reduced stress levels at Week 4 to a sim-
ilar extent (mITT, 37%; PP, 38%) (Table 3) as Mg alone at Week 8 (mITT and PP, 41.3%).
These data suggest Mg–vitamin B6 relieved stress more rapidly during the treatment period
compared with Mg alone, which may indicate a more rapid onset of action associated with
Mg–vitamin B6 than with Mg alone. These results demonstrate that the combination of mag-
nesium and vitamin B6 was 24% more effective in reducing stress compared with magnesium
alone in this subgroup of severely stressed healthy adults. The beneficial effects of Mg–vitamin
Supporting information
S1 Protocol. Redacted study protocol. Redacted study protocol.
(PDF)
S2 Protocol. Protocol publication approval form. Protocol publication approval.
(DOCX)
S1 Infographic. An infographic illustrating the study and its overall findings.
(PDF)
S1 Table. CONSORT checklist. CONSORT checklist.
(DOC)
Acknowledgments
Thanks to all the subjects and their families.
The authors would like to acknowledge Patrick Griffin, MSc, of iMed Comms, Macclesfield,
UK, an Ashfield Company, part of UDG Healthcare plc for medical writing support that was
funded by Sanofi-Aventis Groupe, Gentilly, France, in accordance with Good Publications
Practice (GPP3) guidelines (http://www.ismpp.org/gpp3).
Author Contributions
Conceptualization: Etienne Pouteau.
Formal analysis: Marmar Kabir-Ahmadi.
Investigation: Etienne Pouteau, Lionel Noah, Andre Mazur, Louise Dye, Juliane Hellhammer,
Gisele Pickering, Claude Dubray.
Methodology: Etienne Pouteau.
Project administration: Etienne Pouteau.
Validation: Marmar Kabir-Ahmadi.
Writing – original draft: Etienne Pouteau, Marmar Kabir-Ahmadi, Lionel Noah, Andre
Mazur, Louise Dye, Juliane Hellhammer, Gisele Pickering, Claude Dubray.
Writing – review & editing: Etienne Pouteau, Marmar Kabir-Ahmadi, Lionel Noah, Andre
Mazur, Louise Dye, Juliane Hellhammer, Gisele Pickering, Claude Dubray.
References
1. de Baaij JH, Hoenderop JG, Bindels RJ. Magnesium in man: implications for health and disease. Phy-
siol Rev. 2015; 95: 1–46. https://doi.org/10.1152/physrev.00012.2014 PMID: 25540137
2. Terry J. The major electrolytes: sodium, potassium, and chloride. Journal of intravenous nursing: the
official publication of the Intravenous Nurses Society. 1994; 17: 240–247.
3. Seelig MS. Consequences of magnesium deficiency on the enhancement of stress reactions; preven-
tive and therapeutic implications (a review). J Am Coll Nutr. 1994; 13: 429–446. PMID: 7836621
4. Cuciureanu MD, Vink R. Chapter 19: Magnesium and Stress. In: Vink R, ed. Magnesium in the Central
Nervous System. University of Adelaide Press; 2015.
5. Whyte KF, Addis GJ, Whitesmith R, Reid JL. Adrenergic control of plasma magnesium in man. Clinical
Science. 1987; 72: 135–138. PMID: 3542342
6. Murck H. Magnesium and affective disorders. Nutr Neurosci. 2002; 5: 375–389. https://doi.org/10.1080/
1028415021000039194 PMID: 12509067
7. Jung K-I, Ock SM, Chung JH, Song CH. Associations of Serum Ca and Mg Levels with Mental Health in
Adult Women Without Psychiatric Disorders. Biol Trace Elem Res. 2010; 133: 153. https://doi.org/10.
1007/s12011-009-8421-y PMID: 19543697
8. Verster JC, Palagini L, Mackus M, van de Loo AJ, Garssen J, Roth T. Insomnia and stress: associations
with mental resilience and mood, Sleep. 2018; A162: https://doi.org/10.1093/sleep/zsy061.427
9. Abbasi B, Kimiagar M, Sadeghniiat K, Shirazi MM, Hedayati M, Rashidkhani B. The effect of magne-
sium supplementation on primary insomnia in elderly: A double-blind placebo-controlled clinical trial. J
Res Med Sci. 2012; 17: 1161–1169. PMID: 23853635
10. Zogovic D, Pesic V, Dmitrasinovic G, Dajak M, Plecas B, Batinic B, et al. Pituitary-gonadal, pituitary-
adrenocortical hormones and IL-6 levels following long-term magnesium supplementation in male stu-
dents. J Med Biochem. 2014: 291–298.
11. McCarty MF. High-dose pyridoxine as an ’anti-stress’ strategy. Med Hypotheses. 2000; 54: 803–807.
https://doi.org/10.1054/mehy.1999.0955 PMID: 10859691
12. Iezhitsa IN, Spasov AA, Kharitonova MV, Kravchenko MS. Effect of magnesium chloride on psychomo-
tor activity, emotional status, and acute behavioural responses to clonidine, d-amphetamine, arecoline,
nicotine, apomorphine, and L-5-hydroxytryptophan. Nutr Neurosci. 2011; 14: 10–24. https://doi.org/10.
1179/174313211X12966635733277 PMID: 21535917
13. Majumdar P, Boylan LM. Alteration of tissue magnesium levels in rats by dietary vitamin B6 supplemen-
tation. Int J Vitam Nutr Res. 1989; 59: 300–303. PMID: 2599796
14. Franchimont, MAG 10 trial, 1990. Non-published data on file.
15. Abraham GE, Schwartz UD, Lubran MM. Effect of vitamin B-6 on plasma and red blood cell magnesium
levels in premenopausal women. Ann Clin Lab Sci. 1981; 11: 333–336. PMID: 7271227
16. Schuchardt JP, Hahn A. Intestinal Absorption and Factors Influencing Bioavailability of Magnesium–An
Update. Curr Nutr Food Sci. 2017; 13: 260–278. https://doi.org/10.2174/
1573401313666170427162740 PMID: 29123461
17. Boyle NB, Lawton C, Dye L. The Effects of Magnesium Supplementation on Subjective Anxiety and
Stress-A Systematic Review. Nutrients. 2017; 9: https://doi.org/10.3390/nu9050429 PMID: 28445426
18. Sanofi. Magne B6 Summary of Product Characteristics. [accessed June 2018]. Available from: http://
agence-prd.ansm.sante.fr/php/ecodex/rcp/R0271599.htm.
19. Brown TA, Chorpita BF, Korotitsch W, Barlow DH. Psychometric properties of the Depression Anxiety
Stress Scales (DASS) in clinical samples. Behav Res Ther. 1997; 35: 79–89. PMID: 9009048
20. Lovibond SH, Lovibond PF (1995). Manual for the Depression Anxiety Stress Scales. ( 2nd. Ed.) Syd-
ney: Psychology Foundation.
21. Elin RJ. Assessment of magnesium status for diagnosis and therapy. Magnes Res. 2010; 23: 194–8.
22. Ruwaard J, Lange A, Bouwman M, Broeksteeg J, Schrieken B. E-mailed standardized cognitive beha-
vioural treatment of work-related stress: a randomized controlled trial. Cogn Behav Ther. 2007; 36:
179–192. https://doi.org/10.1080/16506070701381863 PMID: 17852171
23. Comittee for Proprietary Medicinal Products (CHMP). Points to consider on adjustment for baseline
covariates. 2003. http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/
09/WC500003639.pdf. Accessed April 2018.
24. Dietary Reference Intake. [accessed ] Available from: http://www.nationalacademies.org/hmd/~/media/
Files/Activity%20Files/Nutrition/DRI-Tables/2_%20RDA%20and%20AI%20Values_Vitamin%20and%
20Elements.pdf?la=en
25. Fine K, Santa Ana C, Porter J, Fordtran J. Intestinal absorption of magnesium from food and supple-
ments. J Clin Invest. 1991; 88: 396–402. https://doi.org/10.1172/JCI115317 PMID: 1864954
26. Ranade VV, Somberg JC. Bioavailability and Pharmacokinetics of Magnesium After Administration of
Magnesium Salts to Humans. Am J Ther. 2001; 8: 345–357. PMID: 11550076
27. Firoz M, Graber M. Bioavailability of US commercial magnesium preparations. Magnes Res. 2001; 14:
257–262. PMID: 11794633
28. Lindberg J, Zobitz M, Poindexter J, Pak C. Magnesium bioavailability from magnesium citrate and mag-
nesium oxide. J Am Coll Nutr. 1990; 9: 48–55. PMID: 2407766
29. Walker AF, Marakis G, Christie S, Byng M. Mg citrate found more bioavailable than other Mg prepara-
tions in a randomised, double-blind study. Magnes Res. 2003; 16: 183–191. PMID: 14596323
30. Gloster AT, Rhoades HM, Novy D, Klotsche J, Senior A, Kunik M, et al. Psychometric properties of the
Depression Anxiety and Stress Scale-21 in older primary care patients. J Affect Disord. 2008; 110:
248–259. https://doi.org/10.1016/j.jad.2008.01.023 PMID: 18304648
31. Antony MM, Bieling PJ, Cox BJ, Enns MW, Swinson RP. Psychometric Properties of the 42-Item and
21-Item Versions of the Depression Anxiety Stress Scales in Clinical Groups and a Community Sample.
Psychological Assessment. 1998; 10: 176–181.
32. Tanniou J, van der Tweel I, Teerenstra S, Roes KC. Subgroup analyses in confirmatory clinical trials:
time to be specific about their purposes. BMC Med Res Methodol. 2016; 16: 20. https://doi.org/10.1186/
s12874-016-0122-6 PMID: 26891992
33. De Souza MC, Walker AF, Robinson PA, Bolland K. A synergistic effect of a daily supplement for 1
month of 200 mg magnesium plus 50 mg vitamin B6 for the relief of anxiety-related premenstrual symp-
toms: a randomized, double-blind, crossover study. J Womens Health Gend Based Med. 2000; 9: 131–
139. https://doi.org/10.1089/152460900318623 PMID: 10746516
34. Fathizadeh N, Ebrahimi E, Valiani M, Tavakoli N, Yar MH. Evaluating the effect of magnesium and mag-
nesium plus vitamin B6 supplement on the severity of premenstrual syndrome. Iran J Nurs Midwifery
Res. 2010; 15: 401–405. PMID: 22069417
35. Akarachkova ES, Vershinina SV. [The role of magnesium in neuroprotection and neuroplasticity]. Zh
Nevrol Psikhiatr Im S S Korsakova. 2013; 113: 80–83.
36. Ciaccio M, Bellia C. Hyperhomocysteinemia and Cardiovascular Risk: Effect of Vitamin Supplementa-
tion in Risk Reduction. Curr Clin Pharmacol. 2010; 5: 30–36. PMID: 20236081
37. Schneiderman N, Ironson G, Siegel SD. Stress and health: psychological, behavioral, and biological
determinants. Annu Rev Clin Psychol. 2005; 1: 607–628. https://doi.org/10.1146/annurev.clinpsy.1.
102803.144141 PMID: 17716101
38. DeLongis A, Folkman S, Lazarus RS. The impact of daily stress on health and mood: psychological and
social resources as mediators. J Pers Soc Psychol. 1988; 54: 486–495. PMID: 3361420
39. Charles ST, Piazza JR, Mogle J, Sliwinski MJ, Almeida DM. The wear and tear of daily stressors on
mental health. Psychol Sci. 2013; 24: 733–741. https://doi.org/10.1177/0956797612462222 PMID:
23531486
40. Al-Sayed NN, Elsheikh MM, Mahmoud DAM, El-Ghamry RH, Sabry WM, Hatata HM, et al. Work stress:
psychological impact and correlates in a sample of Egyptian medical residents. Middle East Current
Psychiatry. 2016; 23: 113–118.
41. Piazza JR, Charles ST, Sliwinski MJ, Mogle J, Almeida DM. Affective reactivity to daily stressors and
long-term risk of reporting a chronic physical health condition. Ann Behav Med. 2013; 45: 110–120.
https://doi.org/10.1007/s12160-012-9423-0 PMID: 23080393
42. Therabel La Meuse. Magnespasymyl Summary of Product Characteristics.